Introduction

Most currently marketed antibiotics are natural products of microbial origin, and >120 of the most important medicines in use today are obtained from terrestrial microorganisms.1, 2 Often because of drug-resistance phenomena, 17 million lives every year are lost to infectious diseases,3 leading to the global concern that we may soon be facing a post-antibiotic era with reduced capabilities to combat microbes. As a consequence, a concerted worldwide search for new antibiotics from microbial origin is on-going, with focus on the potential of marine and terrestrial bacteria as source for novel metabolites with interesting biological and pharmaceutical properties.4, 5, 6, 7 Diverse habitats, for example, tropical forests, deep sea sediments, sites of extreme temperature, salinity or pH, were explored and were successful to yield new microorganisms, which in turn provide the potential for novel metabolic pathways and new bioactive natural compounds.8 Streptomyces spp. are widespread in nature and continue to have a significant role in the production of bioactive metabolites. Streptomyces spp. produce many classes of secondary metabolites with great biofunctional diversity (antibiotics, antifungal, antiviral, anticancer, immunosuppressants, insecticides, herbicides etc.) and diverse chemical structures, which makes them useful as pharmaceuticals and agricultural agents.1, 9, 10

During our continued search for bioactive constituents from bacteria, strain DGC1 was isolated from a soil sample collected from the Devil’s Golf Course salt pan (Death Valley National Park, CA, USA). Phylogenetic studies of DGC1 strain were conducted as described earlier,11 and the amplified 500-bp 16S rRNA fragment was found to have 99% identity (BLAST) to the 16S rRNA gene of Streptomyces atrovirens strain NRRL B-16357. The extract obtained from the small-scale fermentation of Streptomyces sp. DGC1 on SG medium12, 13 exhibited several unusual green fluorescent bands under long UV (365 nm), which stained to yellow with anisaldehyde/sulfuric acid in the pre-screening. A large-scale fermentation of the strain in SG medium afforded a crude extract from which different chromatographic techniques led to the isolation of five new benzamides: pyramidamycins A-D (2–5) and 3-hydroxyquinoline-2-carboxamide (6), whose structures were determined by NMR (1D and 2D) spectroscopy and MS (ESI and HR-ESI) studies (Figure 1). Benzamides are of increased interest, because Ning et al.14 have demonstrated recently that the synthetic benzamide chidamide is a potent histone deacetylase inhibitor in T-cell lymphoma cell lines. The new compounds were examined for antimicrobial and cytotoxic activities.

Figure 1
figure 1

Chemical structures of compounds 17.

Results and Discussion

In our search for new bioactive compounds from streptomycetes, Streptomyces sp. DGC1 (Supplementary Figure S2) was cultivated on ISP4 agar plates at 28 °C for 3 days. After grown over, small agar pieces (circa 1 cm3) of the strain were used to inoculate 12 2-l Erlenmeyer flasks, each containing 670 ml of SG medium.12, 13 The cultures were kept on a rotary shaker for 4 days at 28 °C. The reddish brown broth was harvested, mixed with celite, filtered off and extracted with ethyl acetate, and the mycelium was extracted with ethyl acetate followed by acetone. The combined organic extracts from supernatant and cells were concentrated in vacuo to afford 2.30 g of yellow-solid crude extract.

A TLC analysis of the strain extract exhibited several UV yellowish-green fluorescent bands at 366 nm, which turned yellow by staining with anisaldehyde/sulfuric-acid spraying reagent. The HPLC-MS analysis of the crude extract displayed several components with UV spectrum (Supporting Information, Supplementary Figures S3 and S4). Work-up and purification of the 2.30-g crude extract using various chromatographic techniques (Figure 2) led to the isolation of five new compounds, including pyramidamycins A-D (25) and 3-hydroxyquinoline-2-carboxamide (6), all five possessing an amide group (-CONH2). In addition, the five known compounds 2-aminobenzamide (anthranilamide, 1, Supplementary Figures S5–S8),15, 16 4′,7-dihydroxyisoflavanone (daidzein, 7, Supplementary Figures S48–S56),17, 18 2′-deoxy-thymidine,19 2′-deoxy-uridine20 and adenosine,19, 21, 22 were also isolated and characterized.

Figure 2
figure 2

Work-up procedure of extracts from Streptomyces sp. DGC1.

Structure elucidation

The physicochemical properties of compounds 16 are summarized in Tables 1 and 2. The known compounds were identified from their NMR and mass data, by comparison with literature data. Structures 1 and 7 were determined by 1D and 2D NMR studies (Supplementary Figures S5–S8, and S48–S53), and by comparison with literature data.

Table 1 Physicochemical properties of pyramidamycins A-C (24)a
Table 2 Physicochemical properties of pyramidamycin D (5) and 3-hydroxyquinoline-2-carboxamide (6)a

Compound 2 was obtained as a white solid. It is UV absorbing, exhibits a blue fluorescence under long UV (365 nm) and gave a pale-yellow color discoloration on spraying with anisaldehyde/sulfuric acid. The molecular formula of 2 was determined by HR-ESI MS as C8H9NO3 (Table 1, Supplementary Figures S9–S22). The proton NMR spectrum of 2 in DMSO-d6 (Table 3) displayed one chelated broad signal for a hydroxyl group at δ 13.40 along with two broad singlets at δ 8.21 and 7.68, typical for an amide group (-CONH2), which converted to a broad signal (as known from anthranilamide 1) of 2H at δ 5.75 when measured in CDCl3 solvent. In addition, the 1H NMR spectrum displayed ortho-coupled protons at δ 7.76 (d, J=9.0 Hz) and 6.43 (dd, J=9.0, 2.5 Hz), a meta-coupled proton at δ 6.39 (d, J=2.5 Hz) as well as a methoxy singlet at δ 3.75 (s), representing a trisubstituted benzene. The 13C NMR/HSQC spectra (Table 4) confirmed compound 2 to be 2-hydroxy-4-methoxybenzamide and showed the OH group at C-2 (δ 163.6) chelated with the amide carbonyl (δ 172.4) and the methoxy group (δ 55.5) located at C-4 (δ 164.0). The HMBC correlations (Figure 3) of compound 2 finalized the structure, showing 3J correlations from the doublet proton H-6 (δ 7.76) to the amide carbonyl (δ 172.4), C-2 (δ 163.6) and C-4 (δ 164.0). The methoxy group (δ 3.75) could be determined as being attached at C-4 due to its significant HMBC correlation with C-4 (δ 164.0). Furthermore, NOESY correlations between this methoxy group and H-3 as well as H-5 were observed, all of which confirmed structure 2 as 2-hydroxy-4-methoxybenzamide, (Figure 3, Tables 3 and 4). A database search (chemical abstracts) confirmed the novelty of structure 2, which was subsequently named pyramidamycin A.

Table 3 1H NMR (500 MHz) data of the benzamides 1–5 in DMSO-d6, δ in p.p.m. relative to TMS, multiplicities (J Hz−1)
Table 4 13C NMR (125 MHz) data of benzamides 15 in DMSO-d6, (δC, mult.)
Figure 3
figure 3

1H-1H-COSY (bold lines) and selected HMBC (→) correlations in compounds 17.

Compound 3 was obtained as a colorless solid, with a molecular weight of 183 Da corresponding to a molecular formula of C8H9NO4, as deduced by HR-ESI MS (Table 1, Supplementary Figures S23–S27). The proton NMR spectrum (Table 3) and the 13C NMR/HSQC spectra (Table 4) of 3 showed that it contains the same benzamide core as compound 2 with Δm/z=16 amu higher than 2 corresponding to an additional oxygen atom. An additional broad signal at δ 9.86 in the proton NMR spectrum and the absence of the meta-coupled aromatic proton (at C-3 of compound 2) suggested that the extra OH group (δ 9.86) might be located at C-3 (like in the hypothetical structure 8, Figure 4). However, based on the full 2D NMR studies, the methoxy group (δ 3.68) showed an HMBC correlation with C-3 (δ 135.1), confirming its linkage at C-3, rendering the OH group at position 4 (structure 3). All of the remaining HMBC correlations (Figure 3) and NMR data (Tables 3 and 4) are in full agreement with structure 3. Compound 3 is a new structural analogue of 2, 2,4-dihydroxy-3-methoxybenzamide, and was named pyramidamycin B.

Figure 4
figure 4

Chemical structures of compounds 810.

Closely related to pyramidamycin B (3) compound 4 was obtained as a white powder from the same fraction FIII, exhibiting a molecular formula of C8H10N2O3 (HR-ESI MS), which is 1 amu smaller than 3, indicating that one of the OH groups was replaced by an NH2 group (Table 1). The 1H and 13C NMR data of 4 were similar to those of 3 (Tables 3 and 4), giving two alternative possible structures (4 and 9, see Figure 4 for alternative structures 810) depending on the positions of the methoxy and the amino groups. In the HMBC spectra, a 3J correlation was observed from the methoxy group (δ 3.79) to C-4 (δ 149.7), confirming its linkage to C-4 as in compound 2, and not to C-3 (δ 124.9) as in 3. All the remaining HMBC correlations (Figure 3) and NMR data (Tables 3 and 4) are in full agreement with structure 4. Therefore, structure 4 was determined as 3-amino-2-hydroxy-4-methoxybenzamide, and consequently named pyramidamycin C (for spectra, see Supplementary Figures S28–S34).

Structurally related to pyramidamycin C (4), compound 5 was obtained as an orange solid, with a molecular formula of C10H12N2O4 (HR-ESI MS); that is by 42 amu (typical for an acetyl group) higher than 4 (for physicochemical properties, see Tables 1 and 2). The comparison of the NMR data of compound 5 with those of pyaramidamycin C (4) confirmed that 5 contains an additional acetyl group (-COCH3), and the chemical shift of its carbonyl (δ 168.3) suggested an amide or ester connectivity, leaving the two alternative structures 5 and 10. The proton NMR spectra of 5 showed a chelated broad signal at δ 13.52, indicating a free hydroxyl group at C-2 and a broad signal of an NH group at δ 8.87, thus excluding the isomeric structure 10. Compound 5 was further subjected to 2D NMR (HSQC and HMBC) experiments, and 2J correlations were observed from the NH (δ 8.87) and CH3 (δ 1.96) to the carbonyl at δ 168.3, confirming the acetamide moiety. The observed 3JC-H HMBC coupling from the methoxy group (δ 3.80) to C-4 (δ 159.4) also confirmed its linkage to C-4. Also the remaining HMBC couplings (Figure 3) and the other NMR data (Tables 3 and 4) are in full agreement with structure 5. Thus, compound 5 was identified as 3-acetamido-2-hydroxy-4-methoxybenzamide, and named pyramidamycin D (for spectra, see Supplementary Figures S35–S39).

Compound 6 was isolated from fraction FII as a pale-yellow solid. It shows green fluorescence under long UV (365 nm), and has a molecular weight of m/z 188, corresponding to the molecular formula C10H8N2O2 determined by HR-ESI MS. The 1H NMR spectrum revealed signals for a disubstituted benzene ring, a chelated OH group (δ 12.32), one singlet aromatic proton (δ 7.75) along with the typical broad signals of the amide protons (-CONH2 ) as in the above discussed compounds (15, Figure 3). The 13C NMR/HMQC spectra revealed ten carbons, five sp2 methine (δ 129.9, 129.6, 128.3, 127.2 and 120.5, Table 5) and four quaternary sp2 carbon atoms (δ 171.8, 153.9, 141.6, 135.8 and 132.2), of which the first one is the carbonyl amide. In the HMBC spectrum (Figure 3), the disubstituted benzene ring was confirmed and the chemical shift of one of its quaternary carbons at δ 141.6 should be linked to a heterocyclic nitrogen atom, such as in indole or quinoline moieties. The remaining singlet methine proton was in peri-position to H-5 (δ 7.82) based on the observed HMBC correlations between H-4 (δ 7.75) and C-5 (δ 127.2) and from H-5 (δ 7.82) to C-3 (δ 120.5). In addition, 3JC-H HMBC correlations (Figure 3) were observed from the amide protons (-CONH2) and the singlet methine H-4 (δ 7.75) to C-2 (δ 135.8), along with the 2JC-H HMBC coupling of H-4 (δ 7.75) to C-3 (δ 153.9), confirming the structure of 6 as 3-hydroxyquinoline-2-carboxamide. Based on literature search, compound 6 is a new natural product. The compound was previously mentioned by Kaneko and Yameda23 as one of the intermediates of their synthesis of 3-hydroxyquinoline derivatives, however, no NMR and MS data were reported (for spectra, see Supplementary Figures S40–S47).

Table 5 1H (500 MHz) and 13C NMR (125 MHz) data of 3-hydroxyquinoline-2-carboxamide (6) in DMSO-d6, δ in p.p.m. relative to TMS

Biological activity

Cytotoxicity assays

The cytotoxic activity of anthranilamide (1), pyramidamycins A-D (25), 3-hydroxy-quinoline-2-carboxamide (6) and isoflavanone 7 was determined in comparison with the known strong cytotoxic angucyclin landomycin A, using PC3 (Prostate cancer) and H460 (non-small cell lung cancer) cell lines (Figures 5a and b, Table 6). Cell-viability assays showed that pyramidamycin C (4) was the most active compound against both PC3 and H460 cell lines (GI50=2.473 and 7.339 μM, respectively). It exhibited cytostatic activity at all tested concentrations, but was less active than landomycin A (GI50=0.5505 and 4.109 μM, respectively). Pyramidamycin C (4) has a free amino group (-NH2) connected at C-3, which may be responsible for its cytostatic activity. Acetylation of this amino group, as found in the congener pyramidamycin D (5), led to the complete loss of cytostatic activity. All other compounds showed low or no activity, except 3-hydroxy-quinoline-2-carboxamide (6), which revealed some moderate cytostatic activity in the PC3 cell line (Figure 5a, Table 6).

Figure 5
figure 5

Dose–response curve of anthranilamide, pyramidamycins A–D, 3-hydroxy-Quinoline-2-carboxamide and isoflavanone in PC3 (a) and H460 (b) cell lines at 48 h.

Table 6 Cytotoxic activity of pyramidamycin C (4), 3-hydroxy-quinoline-2-carboxamide (6) and 4′,7-dihydroxyisoflavanone (7) in comparison with landomycin A (GI50 values, μM)

Antibacterial activity

The antibacterial activity of compounds 17 were also determined against the Gram-negative bacterium Escherichia coli DH5α (Invitrogen, Grand Island, NY, USA) and the Gram-positive bacteria Micrococcus luteus NRRL B-2618 and Kocuria rosea B-1106 (Table 7). K. rosea B-1106 was included in the testing as a second representative Gram-positive bacterium of the family Micrococcaceae. Of these compounds, benzamides 13 were the most active against the Gram-positive bacterium K. rosea B-1106, with 1 producing a diameter halo of 13±2 mm. Compounds 5 and 6 were slightly active against Gram-negative E. coli DH5α (5=10±2 mm, 6=8±2 mm). Furthermore, compound 6 demonstrated the widest range of activities of those compounds tested, inhibiting also the Gram-positive M. luteus NRRL B-2618 (9±2 mm).

Table 7 Diameter halo measurements (ml) of compounds 17 tested against Gram-positive (K. rosea NRRL B-1106 and M. luteus NRRL B-2618) and Gram-negative bacteria (E. coli DH5α) at 100 μg per disc

Experimental Procedure

General experimental procedures

UV spectra were recorded on a Shimadzu UV-1800 (Model TCC-240A) UV spectrometer (Shimadzu, Kyoto, Japan). NMR spectra were measured on a Varian VnmrJ 500 (1H, 500 MHz; 13C, 125.7 MHz) spectrometer; the δ values were referenced to the solvent signals (δ2.5 and 40.6 p.p.m., respectively, for DMSO-d6, and δ7.21 and 77.06 p.p.m., respectively, for CDCl3). ESI mass spectra were recorded on a Finnigan LCQ ion trap mass spectrometer (Thermo Finnigan, Waltham, MA, USA). HR-ESI mass spectra were recorded on an Agilent LC/MSD TOF (resolution: 10 000; 3 p.p.m. mass accuracy; inlet systems: Agilent Technologies 1200 Series LC pumps) mass spectrometer (Agilent Technologies, Palo Alto, CA, USA). Samples were introduced by means of a syringe pump. Prep C18 7-μm column (7.8 × 300 mm2) on a binary LC system (solvent A: H2O/0.2%/formic acid; solvent B: acetonitrile; flow rate: 2.0 ml min−1; 0–15 min, 0–75% A (linear gradient); 15–20 min, 0% A and 100% B; 20–22 min, 0–75% A (linear gradient); 22–27 min, 75% A). HPLC-MS analyses were carried out using a Symmetry Anal C18 5-μm column (4.6 × 250 mm2) on a binary LC system. Flash chromatography was carried out on silica gel MN 60 (140–270 mesh ASTM). Rf values were measured on Polygram SIL G/UV254 (Macherey–Nagel & Co., Düren, Germany). Size-exclusion chromatography was performed on Sephadex LH-20 (GE Healthcare, Little Chalfont, Buckinghamshire, UK).

Taxonomy

A soil sample containing DGC1 was collected from the Devil’s Golf Course salt pan (Death Valley National Park). Approximately 1 g of soil sample was suspended in 25 ml sterile H2O and was subsequently heated to 80 °C water for 30 min. After heating, the sample was briefly vortexed and the large sediment was allowed to settle before plating 100 μl serial dilutions of the supernatant (10−1, 10−2, 10−4 and 10−8) onto oatmeal-agar plates supplemented with nalidixic acid (1 mg) and cycloheximide (10 mg). After 5 days of incubation at 30 °C, several bacterial colonies were observed. After a total incubation time of 7 days, over 50 individual colonies were streaked on oatmeal, M2, ISP224 and ISP424 agar plates. Colonies with general morphological similarities to streptomycetes were taken and streaked onto agar plates, in which they were found to have optimal growth (oatmeal, M2, ISP2 or ISP4). Finally, strains were grown in tryptic soy broth liquid media and stored as glycerol stocks for further study.

Phylogenetic studies of DGC1 were conducted as previously described.11 The amplified 500-bp 16S rRNA fragment was found to have 99% identity (BLAST search) to the 16S rRNA gene of S. atrovirens strain NRRL B-16357. The sequenced 16S rRNA gene fragment from DGC1 has been deposited in the NCBI nucleotide database with an accession number JN836739.

Cell-viability assay

Prostate cancer cell line PC3 and non-small cell lung cancer cell line H460 were used to determine the cytotoxicity of amides 17 in comparison with landomycin A. Experiments were performed in four replicates and conducted as previously described.25

Antibacterial activity test

The Gram-negative bacterium E. coli DH5α (Invitrogen) and the Gram-positive bacteria M. luteus NRRL B-2618 and K. rosea B-1106 were maintained in lysogeny broth liquid media and Mueller–Hinton agar (beef infusion 2 g l−1, casein peptone 17.5 g l−1, starch 1.5 g l−1, Difco granulated agar 17 g l−1, 1 l of double distilled H2O, autoclaved). A sterile loopful of each organism was inoculated into a 7-ml culture of lysogeny broth broth and incubated in a 37 °C orbital shaker at 200 r.p.m. for 10 h. Each test organism was streaked on a sterile Mueller–Hinton agar plate with a sterile cotton swab. Compounds 17 were dissolved in methanol and were aliquoted in 100 μg amounts per each 6-mm sterile filter disc and were allowed to dry in a laminar flow hood. The discs were placed on the plates, which were then incubated for 24 h at 37 °C.26, 27 The resulting diameter halos were measured (Table 7).

Culture material, fermentation and isolation

SG medium

Glucose (20 g, Sigma-Aldrich, Milwaukee, WI, USA), yeast extract (5 g, Acros Organics), soytone (10 g, Becton, Dickinson & Co., Franklin Lakes, NJ, USA), CoCl2 × 6 H2O (1 mg, Acros Organics, Geel, Belgium) and calcium carbonate (2 g, Sigma-Aldrich) were dissolved in 1 l of demineralized water. The suspension (pH 7.2) was sterilized by autoclaving for 33 min at 121 °C.

Oatmeal-agar medium

Oatmeal (60 g) and agar (12.5 g, Becton, Dickinson & Co) were dissolved in 1 l of demineralized water. The pH of the medium was adjusted to 6.0 with 1 M NaOH and 1 M HCl before sterilization.

M2-agar medium

Glucose (4.0 g, Sigma-Aldrich), yeast extract (4.0 g, Acros Organics), malt extract (10.0 g, MP Biomedicals, LLC, Santa Ana, CA, USA) and agar (15.0 g, Becton, Dickinson & Co) were dissolved in 1 l of demineralized water.

Fermentation, extraction and isolation

Streptomyces sp. DGC1 was cultivated on ISP4 agar plates at 28 °C for 3 days. Pieces of well-grown agar subculture of the strain were used to inoculate 12 2-l flasks, each containing 670 ml of SG medium, which was grown at 28 °C and harvested after 4 days. The obtained reddish brown culture broth was mixed with celite and filtered off. The water phase was extracted with EtOAc (4 × 2 l) and the biomass was extracted with EtOAc (4 × 500 ml), then with acetone (1 × 300 ml). The organic extracts were evaporated in vacuo at 38 °C, and combined after evaporation based on the TLC profile affording 2.30 g of yellow-solid crude extract.

Separation of the obtained crude extract on silica gel column (column 2.5 × 50 cm2, 120 g), using a stepwise MeOH/CH2Cl2 gradient (0.4 l 0% MeOH→fraction FI, then 0.2 l 3% MeOH and 0.2 l 5% MeOH combined→fraction FII, then 0.2 l 10% MeOH and 0.2 l 20% MeOH combined→fraction FIII, then 0.3 l 50% MeOH and 0.2 l 100% MeOH combined→fraction FIV), yielded four fractions, FI (350 mg, yellow oil), FII (40.8 mg, yellow solid), FIII (700 mg, yellow solid) and FIV (400 mg, yellow solid). Fraction FI was identified as fats based on TLC, HPLC-MS and anisaldehyde/sulfuric-acid spraying reagent. Fraction FII was further purified using Sephadex LH-20 (2 × 50 cm2, 50% MeOH/CH2Cl2) to give 3-hydroxyquinoline-2-carboxamide (6; pale-yellow solid, 4.2 mg). Purification of fraction FIII was carried out by Sephadex LH-20 (2 × 50 cm2, 50% MeOH/CH2Cl2) followed by HPLC to yield anthranilamide (1; pale-yellow solid, 3.8 mg), pyramidamycin A (2; white solid, 5.6 mg), B (3; white solid, 7.2 mg) and C (4; white powder, 4.7 mg). Finally, fractionation and purification of fraction FIV using Sephadex LH-20, PTLC and HPLC afforded pyramidamycin D (5, 7.2 mg), 4′,7-dihydroxyisoflavanone (7; yellow solid, 5.3 mg), 2′-deoxy-thymidine (white solid, 3.8 mg), 2′-deoxy-uridine (white solid, 5.1 mg) and adenosine (white solid, 10.3 mg) in pure forms (Figure 2).

2-Aminobenzamide; anthranilamide (1). Pale-yellow solid; UV absorbing (254 nm), blue fluorescence under long UV (365 nm); Rf 0.59 (7% MeOH/CH2Cl2), 0.31 (CH2Cl2); yellow coloration with anisaldehyde/sulfuric-acid spraying reagent; UV (MeOH) λmax (log ɛ) 214 (4.14), 256 sh (3.66), 330 (3.53) nm; 1H NMR (DMSO-d6, 500 MHz) and 13C NMR (DMSO-d6, 125 MHz), see Tables 3 and 4; (−)-APCI MS m/z 135 [M-H]; (−)-ESI MS m/z 135 [M-H]; (+)-ESI MS m/z 137 [M+H]+.

Pyramidamycin A (2). Physicochemical properties, see Table 1; 1H NMR (DMSO-d6, 500 MHz) and 13C NMR (DMSO-d6, 125 MHz), see Tables 3 and 4; 1H NMR (CDCl3, 500 MHz) δ 12.48 (1H, br s, 2-OH), 7.25 (1H, d, 9.0, H-6), 6.45 (1H, d, 2.5, H-3), 6.40 (1H, dd, 9.5, 2.5, H-5), 5.75 (2H, br s, NH2), 3.80 (3H, s, 4-OCH3); 13C NMR (CDCl3, 125 MHz), δ 172.5 (1-CO), 165.2 (Cq-4), 164.6 (Cq-2), 127.9 (CH-6), 107.5 (CH-5), 101.7 (CH-3), 55.7 (4-OCH3).

4,7-Dihydroxyisoflavanone; daidzein (7). Yellow solid; Rf 0.43 (7% MeOH/CH2Cl2); UV absorbing (254 nm); 1H NMR (DMSO-d6, 500 MHz) δ 10.84 (1H, br s, 4′-OH), 9.54 (1H, br s, 7-OH), 8.28 (1H, s, H-2), 7.96 (1H, d, 8.5, H-5), 6.93 (1H, dd, 8.5, 2.5, H-6), 6.86 (1H, d, 2.0, H-8), 7.38 (2H, dd, 6.5, 2.5, 2′-H, 2″-H), 6.80 (2H, dd, 7.0, 2.0, 3′-H, 3″-H) p.p.m.; 13C NMR (DMSO-d6, 125 MHz), δ 152.8 (CH-2), 123.5 (Cq-3), 174.7 (CO-4), 116.6 (Cq-4a), 127.3 (CH-5), 113.9 (CH-6), 162.6 (Cq-7), 102.1 (CH-8), 157.4 (Cq-8a), 122.5 (Cq-1′), 130.1 (CH-2′/CH-2″), 115.0 (CH-3′/CH-3″), 157.2 (Cq-4′); (−)-APCI MS m/z 253 [M-H].